Christian Weyer

icon indicating a website link https://www.thinktecture.com/christian-weyer icon indicating an email [email protected]

icon company @thinktecture icon location Germany icon location Christian Weyer is co-founder and CTO of Thinktecture AG. He is an engaged and passionate speaker at various software conferences and events all over the world.

Results 179 comments of Christian Weyer

Feature Request: Support for `phi-3-vision-128k-instruct`

Thx. These are not good enough for our use cases ;-).

Classification with Strategy, Mapping of Contract and image comparison

Are you still working on this project @enoch3712 ?

Integrate with new OpenAI json_schema params

> +1. Looks like support was added for adding the `strict=True` argument for tool calls (#938), but will there be a corresponding `Mode.JSON_STRICT` update using the new "json_schema" support in...

Integrate with new OpenAI json_schema params

> i think structured outputs will be better than strict json Would you elaborate a bit on this? Thanks :-)

Different results with and without Instructor

> https://eugeneyan.com/writing/prompting/ > > Could you help me understand what might be the reason for this? > > Without Instructor: Call == Text With Instructor: Call == Text + Output_Schema...

[llamacpp] - Quality of F16 GGUF is (still) worse than the online demo

Hey @tc-mb! Thanks for the updated instructions. I just tried the linked GGUF above with: `./llama-minicpmv-cli -m "MiniCPM-Llama3-V-2_5/ggml-model-Q4_K_M_for_pr.gguf" --mmproj "MiniCPM-Llama3-V-2_5/mmproj-model-f16_for_pr.gguf" -c 4096 --temp 0.0 --top-p 0.8 --top-k 100 --repeat-penalty 1.05...

[llamacpp] - Quality of F16 GGUF is (still) worse than the online demo

I am using the official llama.cpp release. Latest commit from today.

[llamacpp] - Quality of F16 GGUF is (still) worse than the online demo

Just completely deleted the clone, re-cloned llama.cpp, rebuilt everything. Same result... 🤔

[llamacpp] - Quality of F16 GGUF is (still) worse than the online demo

Are you also running on Apple/MPS/Metal @tc-mb ?

[llamacpp] - Quality of F16 GGUF is (still) worse than the online demo

> well, I still use my fork llamacpp. I will try official llama.cpp now. I use MacBook Pro with M2. Ah, OK. I thought that everything 'good' was already merged...